Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Classification of different level of Aflatoxin B1 on corn kernels surface using short-wave infrared reflectance hyperspectral imaging

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Paper number  131596857,  2013 Kansas City, Missouri, July 21 - July 24, 2013. (doi: @2013
Authors:   Wei Wang, Kurt C. Lawrence, Gerald W. Heitschmidt, William R. Windham, Yankun Peng, Xuan Chu, Nannan Zhang
Keywords:   Aflatoxin B1 Hyperspectral imaging Maize kernel Variables importance in projection PLS-DA.

Abstract. AFB1 has been classified as a class 1 human carcinogen by the International Agency for Research on Cancer. In this paper a shortwave infrared (SWIR) hyperspectral imaging system was used to assess the potential to detect low levels of Aflatoxin B1 (AFB1) contaminants on the surface of healthy corn kernels. Four different AFB1 solutions were prepared and deposited on kernel surfaces to achieve 10, 20, 100, and 500 ppb respectively, and a total of 120 kernels attributed to the four classes were prepared. Control samples were comprised of 30 healthy kernels. The optimal wavelengths that gave the highest contrast between sound and fungal infected corn kernels were selected depending on Variables Importance in Projection (VIP) scores extracted from Partial Least Square (PLS) analysis. Then, a PLS-Discriminant Analysis (PLS-DA) method was used to identify different classes. The result demonstrated that a relative better classification result could be obtained using those wavelengths selected by means of the VIP procedure than that using the overall wavelengths between 1000 and 2500nm. Although the overall classification accuracies (70%) was not good enough now, some important key wavelengths which represent the AFB1, alcohol functional group, and corn components such as starch, protein and cellulose were identified, which indicates the usefulness of the method proposed in this paper.

(Download PDF)    (Export to EndNotes)